AC-FEUP-G75

Data Mining Project To Loan or not to Loan

Project Group - G75

Project Members:

• Hugo Guimarães - up201806490
• Beatriz Mendes - up201806551
• Henrique Pereira - up201806538

Usage

All the available code can be accessed in the src/jupyters

Requirements:

• Python
• Jupyter Notebook

Given the runnable code is done through jupyter notebooks, there is no need to install further dependencies, as the notebooks have been executed and delivered in a way that the desired output of each file can be seen.

The project contents are distributed like this:

• docs
  • Presentation_G75.pdf - Presentation displayed in class at 15/12/2021
  • report_G75.pdf - Final Report PowerPoint as a pdf file for the 23/12/2021 delivery
• src
  • banking_data - Inicial Provided CSVs
  • csvs - CSV files created by our group
    • results - csvs with te last generated predictions
      • final.csv - StratifiedKFold algorithm generated prediction
      • testing_model.csv - train_test_split algorithm generated prediction
    • loan_united_test.csv - final csv after data cleaning and merging containg the test data
    • loan_united_train.csv - final csv after data cleaning and merging containg the train data
  • database - folder with the database related files (used to replace csv files)
  • jupyters - Code for the project
    • algorithms.ipynb - python notebook containg all the pipeline execution off all our clasification algorithms
    • cleaning.py - python code for the data cleaning process
    • clustering.ipynb - python notebook for data clustering generation
    • graphs.ipynb - python notebook for the graph generation and data analysis
    • goals.md - markdown file for the project goals
    • merge.ipynb - python notebook for data merging into a single sqlite3 final table, to which the algorithms can now be applied

Final Submission

• Final report slides can be seen here